Business Intelligence

Business Intelligence (BI) is the overarching term that describes the systems that use business data as a raw material to produce information that can be used to support empirical decision making processes.

In order for BI to have value, it is necessary for it to be:

• Relevant – The information must be able to be understood by the person receiving it, and they must be able to take action on it. If a person cannot act on a report, it is merely noise.
• Timely – It is important that information is received as soon as possible after the event and at an appropriate frequency.
• Accurate – Information that is not accurate can do more harm than good. Appropriate accuracy tolerances must be established for the system being built. Of course this will depend heavily on the cleanliness of the data being used.

Business Intelligence enables a company to leverage information assets that may be currently dormant.

There are 2 main reasons that a company may want to implement a BI system:

• Efficiency – There may be useful information that is available, but the amount of work necessary to produce it is uneconomic. By automating the production of this information in a BI system, the efficiencies can be such that it is possible to use this information regularly.
• Insights – New business insights can be produced by leveraging existing data and producing actionable insights in forms that the company have never seen before. Often, this will take the form of taking information from many source systems and integrating it to produce new information.

Data warehousing

A data warehouse is a system that is used to collect information from many source systems and arrange it in a format suitable for reporting.  Typically, data in source systems is not very easy to use for reporting as the structures are very complex and different for every system.  A data warehouse solves this problem by putting all of the data in one place and putting it in a format that can be used for reporting and analysis in a consistent way, regardless of what system it originally came from.

Click here to see how Decision Resources helped simplify a global publishing house’s Data Warehouse

Dimensional Modelling

A common technique for analysing and presenting data in a useful format is Dimensional Modelling.  If built correctly, these models can provide an intuitive way for people to ask questions about their business.  Although it sounds quite complex, a dimensional model is actually quite intuitive to use.  Data is arranged into what is called a cube.

The cube has 2 main components:

• Measures – These are what we want to measure. They are usually numerical. Examples could be $, units sold or headcount.
• Dimensions – These are how we want to divide up a particular measure.  They define the ways that we can analyse a measure and the types of questions that we could ask of it.  Examples could be time, product or location.

This diagram shows a sample of a Sales $ measure, modelled with 3 common dimensions.